See also: IRC log, previous 2007-06-05
TomB: Welcome to new members
Justin: US Library of Congress (LC).
<RalphS> Justin Thorp
Justin: research into emerging web technologies. Incorporation into web strategies
<RalphS> Clay Redding
Clay: Also US LC. Library Services
department.
... XML background. Office handles vocabs/code lists, so a natural fit
TomB: Describing telecon procedure.
Guus: Proposed accept miuntues of 5th June.
... all action items there.
... no objections
... Possible telecons cancelled in first two weeks of August
TomB: Already added this to scribing document
(duty/on deck).
... shows which telecons will be cancelled.
<TomB> http://www.w3.org/2006/07/SWD/wiki/ScribeDuty
Guus: Face to Face meeting
... plan to have f2f in fall
... look at results so far.
Tomb: first option W3C Tech plenary week.
... summarising results of poll
... looks like tech plenary is better but Daniel and Guus can't make that
... might be able to get a small meeting of six people or so if half the
maybes turn up
<RalphS> current f2f2 poll results
Tomb: but if all came for tech plenary week could be 12.
Guus: Could we split subjects?
... e.g. RDfa in Boston, SKOS in Korea?
... might be piossible to have small SKOS meeting in Korea based on
... responses to far.
... All options suboptimal.
... Guus and Tom to consider and make a proposal
ACTION: Ralph to update SKOS pages to point to UCR [recorded in http://www.w3.org/2007/06/05-swd-minutes.html#action01] [CONTINUES]
ACTION: Alistair to fix wording on skos issues sandbox [recorded in http://www.w3.org/2007/06/05-swd-minutes.html#action03] [CONTINUES]
ACTION: Alistair to propose minimal fix for resolution of issue 33 [recorded in http://www.w3.org/2007/05/22-swd-minutes.html#action09] [CONTINUES]
ACTION: Alistair will look at raising the examples from the issues to test cases [recorded in http://www.w3.org/2007/05/15-swd-minutes.html#action06] [CONTINUES]
ACTION: Jon and Alistair: Move SKOS issues over from Sandbox to Tracker on an ongoing basis [recorded in http://www.w3.org/2007/05/08-swd-minutes.html#action10] [CONTINUES]
Antoine: Do Jon and Alistair need help on this? Could contribute an hour or so.
Alistair: Yes please!
Antoine: Wording issues. Some issues have different names in sandbox and requirements list
Alistair: Didn't just transfer issues verbatim.
Tried to do some editing
... and choose names that were clearer, Is this an issue?
... didn't think about matching req document.
Guus: Antoine take the token.
Alistair: happy for Antoine to do what he sees fit.
ACTION: Guus to include a reference to the original requirements in the resolution proposal [recorded in http://www.w3.org/2007/06/05-swd-minutes.html#action09] [DONE]
ACTION: Guus to point to typical responses that follow guidelines described in telecon [recorded in http://www.w3.org/2007/06/05-swd-minutes.html#action02] [DONE]
<guus>http://lists.w3.org/Archives/Public/public-swd-wg/2007Jun/0033.html
ACTION: Guus to rephrase current proposal, slightly revise naming based on comments [recorded in http://www.w3.org/2007/06/05-swd-minutes.html#action07] [DONE]
<Bern> http://www.w3.org/2006/07/SWD/wiki/SkosDesign/RelationshipsBetweenLabels/ProposalThree
Guus: Added discussion note.discussions between
antoine/daniel terms are words used to refer to skos concepts
... if you use that then people might think skos:term is what one defines
broader/narrower
... on maybe not an ideal term, but alternatives are worse. Label is
confusing, LabelResource too
... complex, word confuses focus. Guus strong preference for skos:Term
Bernard: Why is label confusing? The whole discussion is confusing
Guus: Label comes from RDF terminology. Every
resource/URI can have a label. Original SKOS had a subproperty of rdsf:label.
Now
... two ways of defining the terms that you want to refer to a concept.
... Either a label (literal). Or as a URI that links to other URIs. prefLabel
will then link to a ... literal rather than a label. So confusing.
... Normal in thesauri terms to talk about terms as being items that are used
to refer
... to a notion in a vocabulary
Bernard: Problem is that we're tryng to address many audiences with different preconceptions
<aliman> +1 on what bernard said
Guus: Can the LC people take a look and advise?
Ed: Spent some time reviewing proposal, getting
up to speed.
... glossary is referenced (willpower). In there ideas like concept and term
are defined
... seemed pretty clear what they meant. In general, want a miminal fix
approach.
... doumentation properties can have three different values literal, resource
description or uri
... why can't we use that?
Guus: Original proposal was to remove a
constraint and make spec smaller. Main problem is
... introducing confusion. Tools need to be on the lookout. To visualise a
concept need to be aware if it's
... pointing to a literal or a resource. Second issue is OWL. Won't be able
to work with
... this as OWL required strict separation of object/data properties. Hence
simple extension
... proposal. Minimal extra thing to handle the OWL issue. If you use
prefLabel it's a literal. If
... prefTerm it's a resource.
Ed: Acronym example does a good job of
describing why terms as resoruces would be useful.
... What's the use case where this came up? Who asked for what
functionality?
Guus: Top of document. Link to candidate requirements and use cases
Antoine: Anchors for each requirement are there.
Guus: Main question was whether the distinction between concept and term is confusing?
Ed: Antoine questioning the use of concept in SKOS? Thesauri don't really have concepts. Is that right?
Antoine: Not questioning the way that it's
done. Agree with this. Issue is that during discussion with
... Daniel there's ambiguity between term based and concept based approach.
Would agreee with
... current stance of SKOS.
Ed: unrelated to this tracking discussion on general SW list. URIs for things that aren't resources.
<aliman> maybe we should deal with the semantics of skos:Concept first?
Ed: Idea of addressing concepts is subtle. Core
guide referes to indirection. Some helpful text about Henry VIII.
... If it's just terms that don't have to map to concepts then it would maybe
be easier.
Guus: Point referred to is a broader issue.
Goes to the root of concept-based representaiton in SKOS
... Would be unwilling to open that door unless there are compelling
reasons.
Ed: No, that's fine.
Guus: Want to start closing issues. Interested in Alistair's opinion
Alistair: Favourite position on this is a move
towards original n-ary relations pattern.
... Original proposal by alistair a little contrived. Want miminal solution.
Nice now that the
... extension propoisal is detailed, can figure out the detailed consquences,
e.g. various entailments
... that would follow. Would like to see alternative proposals presented so
that options can be compared
... side by side. As a solution, concerned that simple extension has
consequences that are hard
... to fathom.
Guus: What are those?
Ali: In the example, there's a term with labels
in two languages. What does it mean for a term to
... be labelled in two languages. Isn't a term embedded in a language?
... Could end up with different patterns of use.
Guus: One resolution is to restrict label property to cardinality 1.
Ali: But then what about transliteration or
character sets? More than one way to write down a term
... in a language.
Guus: Same issue in current SKOS with multiple labels.
Antoine: Support Alistair's proposal for one
rdfs:label per term. Multiple wording is
... a situation where you have mutlple terms each with its label, and
relationships between the,
Alistair: Commenting on Ed's question re. use
cases. Most compelling use cases
... are multilinguality. E.g. Japanese and making links between different
character sets.
... Most important motiviation
Guus: Keen to close. Propose action on Guus to
update w.r.t. Alistair's issues. If you
... want a proposal there, need to present it
Alistair: It's a difficult issue.
Guus: If we resolve an issue and there are technical problems, we can reopen
Alistair: Other issues might be lower hanging
fruit.
... trying to find time to do this.
Guus: Need to tackle the hard issues.
Alistair: When SKOS originally set up. Three
namespaces. SKOS, Mappings and Extensions
... Extensions never stabilised. To move forward, maybe have proposals in an
extensions namespace
... that's published as a note, outside normative recommendation
Guus: Experience of dialects not always
positive. Could put everything into an
... extensions space and then decide on status later.
... Two things. 1/ using extensions space for working things out
... 2/ what is normative.
... Could decide everything in the extensions space is normative.
Alistair: Another issue. Example that arises
when modeliing terms. Simple extension proposal
... prefTerm is functional. So can only be one.
... what happens to entities in an open world? People have always modelled
vocabs
... as data structures within defined boundaries. What if someone comes along
and asserts identity
... between two terms?
ACTION: Guus to update proposal with issue of single vs. multiple labels for terms [recorded in http://www.w3.org/2007/06/12-swd-minutes.html#action09]
ACTION: Alistair to provide details of alternative proposal [recorded in http://www.w3.org/2007/06/12-swd-minutes.html#action10]
ACTION: Antoine to put SimpleExtensionProposal into wiki [recorded in http://www.w3.org/2007/06/05-swd-minutes.html#action08] [DONE]
<TomB> aliman, where is the extension namespace documented?
<aliman> SKOS extensions specification
Guus: Would like to leave SKOS concepts and
ontology classes for now
... Grouping construct no progress. But no actions
Guus: Who owns the issue?
Alistair: Still action on me to propose minimal fix
ACTION: Alistair to propose resolution for Issue-33 (minimal fix) [recorded in http://www.w3.org/2007/06/12-swd-minutes.html#action12]
Guus: Report on SKOS validator. Messages from Antoine and Alistair
Antoine: Recently converted vocab to SKOS.
<aliman> SKOS validation service (alpha)
Antoine: used tool for vocab validation. Went
ok.
... devoted to current version. Might have to be adapted, but it's useful.
Alistair: interesting thing about validation
service is that it bears on how
... how we define broader/narrower/related. Service looks for circularity.
Assumes that circularity is bad.
<RalphS> [I will try to remember to write my comments about naming into email -- I wanted to say that given a choice between choosing terminology that may confuse RDF/OWL folk and terminology that is confusing to the expected SKOS community, I'd make the RDF/OWL folk work harder; we're accustomed to the idea of terminology namespaces already.]
Alistair: circularity is non-sensical. Should
that be embedded in semantics?
... Also not just cycles, but overlap, e.g. if there are multiple paths in
the hierarchy.
Guus: OWL Validator. Gives errors and warnings
Alistair: Yes. Errors and warnings. Currently set up for cycles to be errors
Sean:The question is really what is considered
to be an error and what is considered a warning
... are multiple pathways an error?
Guus: SKOS validator can act as a tool for
candidate req stage later.
... would be useful to look for other tools to support skos.
Sean:Do we have a reasonable definition of "a tool that supports SKOS"?
Alistair: Paper for DC conference last year.
Defined three types of tools
... vocab development
... indexing applications. Read vocab allow tagging
... retrieval applciations. Use the structure and the idnex/metadata. Support
browse/search
... Three classes of application. Look at requirements document. USe cases go
outside of that
... scope.
<RalphS> [I note that some attempt to define what it means to "support SKOS" will be very useful when the WG gets to proposing exit criteria for Candidate Recommendation]
Sean: We should be clear about what we think
implementation will be
... is bound up in test cases too
... a set of concrete test cases could define what it means to support
SKOS
<aliman> SKOS implementations of vocabulary development applications: ThManager and the NSDL Metadata Registry
<aliman> -> http://isegserv.itd.rl.ac.uk/public/skos/press/dc2006/camera-ready-paper.pdf SKOS: Requirements for standardisation (paper for DC2006 which talks about classes of application using SKOS) see also presentation: http://dc2006.ucol.mx/papers/miercoles/10.30/presentation.pdf
ACTION: Ralph propose resolution to Recipe issue 1.3 [recorded in http://www.w3.org/2007/04/24-swd-minutes.html#action09] [CONTINUES]
Elisa: Possibility to take topics one by one
... if we have enough bandwidth on call, take opinion on issues.
Elisa: Keep it tight and make reference to things that people are doing.
Elisa: get people to write one or two
sentences. Which topics to cover in each section
... Perhaps one section per call.
Guus: Happy to put that on the agenda for next week.
Elisa: First section. Named terms using URI referencec
Guus: Link?
Elisa: yes in the agenda
TomB: Given resource constraints, only
realistic thing is to
... try and finish it in a form close to the current one.
<RalphS> Elisa: see "Name Terms using URI References" in -> http://www.w3.org/2006/07/SWD/wiki/VocabMgtDraft Vocab Wiki draft
TomB: five high level points in a page or so
each.
... would still take an awful lot of work. Discussion with Elisa, would be
doable to
... do a high level note, but requires real resource. E.g. SW list discussing
exactly this problem
Guus: We are out of time. Let's take this up next week.
Guus: Ben not here, we'll ask for an update next week.
[adjourned]